Next generation DNA sequencing (NGS) has the power to revolutionize biology and medicine. Determining the sequence of the 3 billion nucleotides in the human genome is no longer a daunting task. However, a major impediment to the full implementation of this technology in the area of cancer detection and research is its high error rate. We have established a method that takes advantage of the complementarity inherent in double-stranded DNA. As a result of sequencing both strands of DNA we can identify true mutations as those present at the same position in both strands. Using this technology we can disregard substitutions due to PCR- amplification since they would only be present in one of the two strands. In addition, we can eliminate most sequencing errors resulting from damage to DNA templates. As a result we have establish a method that is 1000-fold more accurate than standard methods used in next generation DNA sequencing The goal of this application is to validate and further develop the method of Duplex Sequencing. Experiments will be carried out to optimize the recovery of duplex sequences and to probe the exceptional sensitivity of the methodology. To evaluate the utility of duplex DNA sequencing we will analyze hard to sequence segments of the human genome that have been resistant to sequence with accuracy using other approaches. We will also address the question of whether mutations rendering cells resistant to chemotherapeutic agents are preexistent in human cancers prior to the initiation of chemotherapy.